Multi-View 3D Parametric Modeling for Cervical Ultrasound Imaging – A Validation Study

The assessment of cervical spine disease mostly relies on radiography which can cause hazard ionizing radiation to patients. Three-dimensional (3D) ultrasound (US) imaging has emerged as a promising technique for scoliosis diagnosis due to its low-cost, non-radiation and flexible operation. However, research on ultrasound cervical imaging remains limited. The objectives of this paper are to develop a novel Multi-View 3D ultrasound parametric modeling imaging method for cervical spine and to perform a phantom study for evaluating the precision and accuracy. US volumes from posterior, left posterolated and right posterolated views, are registered to the same coordinate system, then fused into a full-view volume. The posture information of each vertebra was derived from the full-view volume and applied to a template model to generate a 3D model of the target cervical spine phantom. The generated cervical spine model exhibited high accuracy and precision and showed a good visualization result, which demonstrated the poential of 3D US imaging for cervical spine disease diagnostics..